WO2016156279A1 - Method for assembling substrates - Google Patents

Abstract

The present invention concerns a method for assembling a first face (3) of a first substrate (1) and a first face (7) of a second substrate (5), characterised in that it comprises: forming an interlayer (11) between the first face (3) of the first substrate (1) and the first face (3) of the second substrate (5), the interlayer (11) being in contact with the first face (3) of the first substrate (1) and with the first face (7) of the second substrate (5); eliminating the interlayer (11) and bringing the first face (3) of the first substrate (1) into contact with the first face (7) of the second substrate (5).

Description

 "Substrate assembly process"

FIELD OF THE INVENTION

 The present invention relates to a method of assembling a first face of a first substrate and a first face of a second substrate. It also relates to a microelectronic system with first and second substrates, these substrates can be assembled using the method.

One field of the invention is 3D three-dimensional integration for electronic devices and in particular microelectronic devices, the term microelectronics including nanotechnologies. More specifically, this invention is particularly applicable to the direct bonding of substrates, for example silicon having surface microelectronic circuits. At least some of these microelectronic circuits can lead to the surface of the substrates and a number will be connected between the two bonded surfaces to obtain an electrical connection through the bonding interface.

 The invention can be used for the manufacture of CMOS imagers of "back-side imagers (BSI)" type, also called rear-panel imagers, and in general, for 3D integration with stacks of CMOS transistors and / or memories.

BACKGROUND

 The race for miniaturization has physical limitations that challenge the planar approach used to this day. Indeed, the small dimensions of the new technological nodes reveal previously insignificant parasitic effects which limit their miniaturization.

 Among the processes under study is a three-dimensional architecture called "3D integration", which consists in vertically stacking electronic components, superimposing chips formed for example on substrates (or plates) on each other and establishing short vertical electrical connections between these components, directly through the different layers. The concept of 3D integration allows devices to be realized with a multitude of functionalities as well as drastically reducing the lengths of the interconnections, thus increasing the communication speeds between the different components.

 3D integration, on the other hand, requires the control of the electrical connections between the different chips stacked vertically. The step of aligning the substrates to be stacked is a crucial step in this regard to obtain high densities of interconnections.

It is known to use an alignment machine to align two substrates. It is then possible to obtain alignments around the micrometer between two substrates 200mm or 300mm in diameter. In this context, several techniques are possible. Microscopes can be inserted between the two substrates which are positioned facing each other (their faces to be assembled extending in parallel) and, with alignment marks, the two substrates are aligned in the two directions of the plane of their faces to assemble. Then the microscopes are removed and the substrates are brought together until contact is obtained which triggers direct sticking. Obviously, this approximation is difficult to control and some misalignment may appear during this movement, which generally limits the alignment to + -5μηη (microns). The substrates are then glued and can no longer be moved relative to each other in the plane of the interface unless provoking a detachment by inserting a wedge at the bonding interface. This corner may allow the plates to take off but, once taken off, the alignment is lost and the alignment must be started again.

 In the case of infrared-transparent substrates (which is the case of non-heavily doped silicon), it is possible to use an infra-red visualization to align the two substrates in the plane. No need here to insert microscopes between the plates. The fact of not having to bring the plates together a great distance after the alignment makes it possible to very significantly limit the potential misalignment induced by the approaching movement. After the collage, thanks to the infra-red, it is possible to check the alignment but if it is not good, as the direct bonding took place, you have to take off the plates to start the alignment again hoping get a better result.

 It is also possible to use a technique that uses microscopes in the visible to align the substrates but do not ask to bring them closer together. Indeed, it is possible to start by bringing very close (about 100 μιτι) the substrates of one another. Then, the substrates are shifted one after the other in the plane to visualize the location of the alignment marks which allows to align the substrates with respect to each other. The final approximation remaining is only about 100 μιτι. With this technique, it is possible to obtain an alignment at + -1 μιτι or even + -0.5μηη.

 These different techniques make it possible to align two substrates but with a still more perfectible precision. It is possible to check the alignment obtained but it is necessary to take off the substrates to start the alignment if it is not satisfactory.

The present invention makes it possible to meet at least certain limitations of the prior art. SUMMARY OF THE INVENTION

 One aspect of the invention relates to a method of assembling a first face of a first substrate and a first face of a second substrate. Advantageously, it comprises:

 a formation of an intermediate layer between the first face of the first substrate and the first face of the second substrate, the intermediate layer being in contact with the first face of the first substrate and the first face of the second substrate;

 an elimination of the intermediate layer and, as a result, bringing into contact the first face of the first substrate and the first face of the second substrate.

 The invention also relates in other aspects to a microelectronic system comprising a first substrate and a second substrate, a first face of each of the first and second substrates facing each other, and an intermediate layer between the first face of the first substrate and the first face of the second substrate, the intermediate layer being in contact with the first face of the first substrate and the first face of the second substrate, the intermediate layer being made of a material in liquid form preferably having a viscosity of less than 100 pascal . second at 20 ° C.

A technical effect produced by the present invention is to allow the approximation with a very small difference, for example of less than 0.1 μιτι, of the two faces through which the assembly is to be performed so as to improve the final alignment of the substrates. , therefore potentially to reduce the size of the electrical members to be connected between the two substrates to increase the density of electronic components and / or make the process of making the two substrates electrically viable. In addition, and regardless of the gap between the substrates, the intermediate layer advantageously forms a fluidic cushion of relative displacement of the two substrates in a plane parallel to the two faces to be assembled. This displacement makes it possible to adjust the alignment as well as possible. BRIEF INTRODUCTION OF FIGURES

 The objects, objects, as well as the features and advantages of the invention will become more apparent from the detailed description of an embodiment thereof, which is illustrated by the following accompanying drawings, in which:

 - Figure 1 illustrates, in section, an embodiment of a first substrate used according to the invention, with a drop of a liquid capable of forming a spacer layer.

 - Figure 2 shows an example, in section, of second substrate.

 FIG. 3 gives a view from below, by its first face, of the second substrate of FIG.

 FIG. 4 illustrates a step of the process of the invention with the formation of an interlayer.

 - Figure 5 shows a step following that of Figure 4.

 FIG. 6 illustrates a procedure for controlling the alignment of the two substrates thanks to the invention.

The drawings are given by way of examples and are not limiting of the invention. They constitute schematic representations of principle intended to facilitate the understanding of the invention and are not necessarily at the scale of practical applications. In particular, the relative thicknesses of the different layers and substrates may not be representative of reality. This is the case for brands allowing the alignment of substrates or cavities apparent at the substrate or substrates.

DETAILED DESCRIPTION

 Before beginning a detailed review of embodiments of the invention, are set forth below optional features that can optionally be used in any association or alternatively.

Depending on these optional features, the process can be: the material of the interlayer has a viscosity of less than 100 pascals. second at 20 ° C, preferably less than 1 pascal. second at 20 ° C;

 at least one of the first substrate and the second substrate comprises at its first face a cavity opening on said first face and on a wafer portion of the substrate.

 - The formation of the intermediate layer may also consist of the formation of a film obtained for example by centrifugation.

 - it is also possible to put several drops. The interlayer material does not need to be evenly distributed.

 a first microelectronic component is formed on the first substrate and a second component on the second substrate, the contacting of the first face of the first substrate and the first face of the second substrate being configured to bring the first and second components into contact with each other; microelectronics.

 The invention relates to the assembly of substrates which are understood here in particular to electronic devices, preferably microelectronic devices, in the form of plates (more commonly known as "wafer", in English), preferably comprising at least one chip and / or or an electronic member, for example an electrically conductive element, for interconnection in particular. However, a substrate may not comprise an active or passive electronic component or component; it may for example be a handle serving, transiently, the transfer of a substrate to another. The first substrate is preferably in the form of a plate, but may also be in the form of a panel. The first substrate may comprise a plurality of layers. Preferably, the substrates are based on one or more non-conductive or poorly conductive materials.

The two substrates used are not necessarily of the same size. For example, one may be a plate comprising a plurality of chips and the other a single chip, so as to perform a report of the type chip on plate (or "chip on wafer" in English). It is understood that the invention can make it possible to stack more than two substrates. They are for example formed of a silicon-based material (for example, polycrystalline silicon, or monocrystalline silicon, or silicon on insulator). The present invention can also be applied to substrates which are not made of silicon but of III-V material such as ΓΙηΡ, AsGa or GaN, or of material such as Germanium, sapphire, silica, glass or aluminum. SiC.

 One or both substrates may comprise one or more semiconducting base layers (in silicon in particular), covered or not with one or more thermal oxide layers, possibly having undergone stages of formation of electronic organs such as CMOS transistors. Electrically conductive parts such as copper zones may also extend along the plane of one side of the substrate, between non-electrically conductive areas, depending on the thickness of the substrate (for example with vias opening on the surface of the substrate). face and connect to electrical parts of the other substrate). The drawings do not represent the possible integrated members on one and / or the other of the substrates.

 FIG. 1 illustrates a first substrate 1 having a first face 3 through which it is desired to assemble the first substrate 1 with a second substrate 5. Preferably, the first face 3 is directed along a main plane (here corresponding to the orientation defined by the xy plane of Figure 6) which does not exclude that the first face 3 has protrusions in salient or hollow form. This is also the case with the presence of marks 2a, 2b serving as alignment marks. The number of brands is not limited. Their shape and depth either. Typically, it will be possible to form two diametrically opposite marks near the wafer 4 of the first substrate 1. Other brands, for example secondary to different locations on the first side, may also be present. In addition, marks 2a, 2b could equally well, in addition or alternatively to the case shown, be located on the face of the first substrate 1 opposite the first face 3.

FIG. 2 illustrates an embodiment of the second substrate 5. It may or may not also have electronic components, some of which may be visible at a first face 7 of the second substrate 5, this first face 7 being intended to be assembled with the first face 3 of the first substrate 1. The second substrate 5 preferably comprises second marks 6a, 6a which may be made identically or differently than those of the first substrate 1. The latter may globally be identical or similar to the first substrate 1. Advantageously, the first face of the second substrate 5 is oriented along a main plane (corresponding to the orientation defined by the xy plane of FIG. 6) adapted to be brought into contact with the first face 3 of the first substrate 1, these two faces 3, 7 being thus parallel.

 Particularly advantageously, the marks 2a, 2b of the first substrate 1 have locations which coincide ideally, in pairs, with the marks 6a, 6b of the second substrate 5 when the first faces 3, 7 of the first and second substrates 1, 3 are assembled.

 According to a preferred embodiment, the marks are produced by a lithography step, followed by an etching step.

 In the case shown, without limitation, the second substrate 5 has cavities 9 recessed relative to the face 7. A single cavity could be formed. The cavity or cavities 9 advantageously extend along the first face 7 so as to open on at least one edge (preferably two edges) corresponding to the wafer 8 of the second substrate 5. Thus, each cavity 9 opens at the same time on the first face 7 and on the edge 8 of the second substrate 5. As an indication, the illustration of FIG. 3 shows cavities extending rectilinearly. They can be parallel to each other. Indicative dimensions are Ι ΟΌμιτι wide for 50μηη depth (along the z axis). Of course the design of the cavities may be different and more complex to take into account the electronic components present on each slice. For example, cavities can make turns to bypass components.

Conventionally, the substrates can be maintained by holding devices in the form of supports generally called "chuck" in English; the supports allow in particular the relative displacement of the substrates in the xy plane. They can also participate in the approximation of substrates along the z direction. These devices allow for example a applying the face of a substrate 1, 5 opposite to the first face 3, 7, by a suction circuit. If suction is stopped, hold is disabled.

 The invention allows the relative good positioning and the assembly of the two substrates thus described. One aspect of the invention consists in inserting, at the bonding interface of the substrates, a material allowing the substrates to be very closely brought together (possibly less than 0.1 m) and, preferably, then allowing align the substrates or optimize a first alignment. An interlayer 1 1 is thus formed with this material in contact with the two faces 3, 7. It is the thickness of this layer of material which will define the separation distance between the first and second substrates 1, 5. inserted material is such that it will then disappear to allow the assembly, preferably by direct bonding, to unfold. This elimination does not necessarily imply a complete elimination of the material of the layer 1 1. Indeed, and particularly when it comes to water, especially during a hydrophilic bonding, this material may remain residual, particularly in the bottom of asperities on the surface of the faces of the substrates. The elimination of the interlayer 1 1 therefore means that the deletion of the layer 1 1 itself, but residues from this layer does not prevent the contacting of the two faces 3, 7 are not not excluded, these residues no longer forming a layer.

Figure 4 illustrates the formation of the layer 1 1 maintaining the two substrates 1, 5 spaced apart. The thickness of the layer 1 1 can be very small, for example less than 100 nm, or even less than 10 nm. The layer 1 1 preferably covers the entirety of the first face 3 and the first face 7 or, if the substrates are of different sizes, the entire interface between the two substrates, namely their assembly area . Alternatively, the layer 1 1 may be discontinuous in that it comprises several portions not completely covering the surface of the interface between the two substrates. For example, this embodiment can be produced when the material of the interlayer 1 1 is deposited in the form of several spaced drops on the surface of one of the substrates; it is possible in this case that the spreading of the drops does not produce a junction of the zones of the material thus deposited and / or a complete overlap of the bonding interface of the two substrates. Different materials can be used for this layer 1 1. Although solid phase materials under the process conditions are not excluded because they allow to maintain the desired space between the substrates, it is clearly preferred to use a fluid material, preferably in the liquid phase. Indeed, it is then possible to play on the viscosity of the material to move the substrates 1, 5 and adjust their alignment, using the marks 2a, 2b, 6a, 6b as benchmarks. The shear stress that must be applied to enable displacement is advantageously quite low and the viscosity of the material is therefore preferably less than 100 Pa · s. In the case of a visualization of the marks by optical transmission through the substrates, in particular by infrared rays or by conventional microscopy, for example having taken care to postpone the alignment marks on the rear face of the substrates or using transparent substrates, the material of the layer 1 1 will preferably be transparent respectively to infra-red or in the visible. The material may be, for example, water, alcohol, or a solvent such as acetone. It may have a low angle of drop, for example less than 10 ° vis-à-vis the surface of at least one of the first and second faces 3, 7. It may also be non-wetting and have a large drop angle , for example greater than 60 ° with respect to at least one of the first and second faces 3, 7, in order to modify the interaction with said surfaces and to optimize their parallelism. It is further desirable that the material employed is incompressible.

 During the alignment adjustment phase, the capillarity may be sufficient to hold in place the interlayer 1 1. It is also possible to adjust the hydrophilic nature of the liquid relative to the surface of the first substrate 1 to optimize this retention.

Once the correct alignment is obtained, if the material is sufficiently viscous so that the alignment does not change leaving the substrates 1, 5 free with respect to each other, it is then possible not to maintain the substrates during that the interfacial material disappears. In the opposite case, it is preferable to maintain the alignment of substrates 1, 5 while the interfacial material disappears. To remove the material of this layer 1 1, it can be allowed to evaporate at room temperature and pressure if its vapor pressure is sufficient. It is possible to increase this evaporation by reducing the total pressure around the substrates or by simply reducing the partial pressure of the material (an anhydrous atmosphere in the case where the interlayer material is water for example). It is also possible to increase the temperature. In order to facilitate the disappearance of this material it is also possible to use cavities 9 previously described at the bonding interface and opening on the edge of the plates. In the case of the figures, only the substrate 5 comprises cavities 9 but in addition or alternatively, the first substrate 1 could comprise.

 It may be possible to correct the alignment as the material disappears which corrects the alignment while the substrates 1, 5 are closer and closer to each other.

 The evacuation velocity of the material of the intermediate layer can be adjusted according to, in particular, the number of cavities. For example, there may be a vaporization of the material and this vaporization will be all the more important that the cavity surface will be high.

 As soon as the layer of material has disappeared preferably without residue of its material, the assembly can be carried out, preferably by direct bonding. The substrates 1, 5 are then glued, thus sealing the alignment. According to a preferred embodiment, the joining step is performed from the first faces of the first and second substrates. Bonding mechanically secures and advantageously also produces electrical continuity between at least one member integrated in the first substrate 1 and another member integrated in the second substrate 5.

The direct bonding method allows direct bonding between substrates, requiring neither bonding material nor the need for significant heating or high pressure. This technology is based on molecular adhesion phenomena between the atoms of the two opposite surfaces. The bonding step is preferably followed by annealing. The annealing has the advantage of reinforcing the bonding forces between the first substrate 1 and the second substrate 5.

 Since the plate alignment control can be done at a very short distance and as the invention allows to maintain this alignment as the plates get closer, a very high accuracy can be achieved. It is thus possible to obtain misalignments of less than 100 nm.

 Figure 5 illustrates the result of the bonding. The faces 3, 7 are in contact and secured, the marks being well aligned thanks to the invention.

 Figure 6 shows an example of alignment control prior to gluing. The relative positioning step preferably comprises a step of matching the marks of the first substrate with the marks of the second substrate. Means in particular the cooperation of the projection, in a plane perpendicular to the thickness of the substrate, a mark of the first substrate with the projection, in a plane perpendicular to the thickness of the substrate, a mark the second substrate so that said projections co-operate by overlapping or interleaving.

Particularly advantageously, a detection of the optical transmission mapping is performed through the superposed substrates. By optical transmission is meant the transmission of information transported by light waves through the substrates, at the level of the marks. An optical transmission system has, for checking the alignment of two marks, two components: a light source (the optical transmitters 13a, 13b shown diagrammatically in FIG. 6) and a light detector (the optical receivers 12a, 12b shown diagrammatically in FIG. ). The optical transmission and reception means are positioned so as to emit and respectively receive a signal perpendicular to the thickness of the first substrate 1 and through the thickness of the second substrate 2, and also through the thickness of the interlayer 1 1. In this embodiment, the optical transmission and reception means are positioned so as to emit and respectively receive a signal perpendicular to the thickness of the substrates, ie along the z axis of FIG. 6. Two embodiments of the method of the invention are given below. The characteristics of these examples can be implemented separately and the aspects of two examples can be the subject of any combination between them to achieve the method of the invention.

 In a first example, two silicon-based substrates, in particular 200 mm in diameter, having a resistivity p of, for example, between 1 and 50 ohm / cm and preferably having alignment marks on the front face (the faces 3 and 7 ) are cleaned, for example, with a solution of ozonated water containing 40 mg / l of ozone, with a solution of APM (ammonium peroxide mixture) with a concentration of ammonia, hydrogen peroxide, and deionized water respectively of 0.25 / 1/5 in volume proportions. Then the two substrates 1, 5 are dried in a dryer.

To form the interlayer 1 1, a drop (for example of water), which may be of a volume of 3.2 mm 3 for a first face of 200 mm, is then deposited in the center of the substrate below advantageously in a position such that the first face is horizontal. This is shown in Figure 1 with a drop 10. This drop 10 may be single or not. It is preferably placed in the middle of the first face 3 of the first substrate 1. Then the substrates are brought together to spread the drop from the center and leave a liquid mattress with a thickness possibly of 100 nm between the substrates. This approximation can be completely mechanical by at least one holding device, or simply by dropping, by gravity, the second substrate 5 on the first. It can advantageously be a combination of the two to spread the drop well from the center of the substrate 1 towards the edges and then allow the gravitation to finish the approximation using the liquid mattress that is formed to ensure the physical separation of the substrates before the end of alignment. This alignment can take place during the approximation and after the latter or only once the faces 3, 7 are very close to each other. The alignment control can be done as in Figure 6 by infra-red. Once the alignment is completed, the vacuum can be made around the substrates 1, 5 to remove the interlayer 1 1. As long as the material of the layer 1 1 is not removed and the direct bonding has not been engaged, the alignment is checked and corrected if necessary. As soon as Direct bonding is triggered, alignment control is no longer needed. It is desirable to keep the substrates under vacuum until the direct bonding is effective over the entire surface. An alternative may be to replace the vacuum with an anhydrous nitrogen atmosphere having a dew point temperature of less than -10 ° C or less than -85 ° C.

 According to a second example, a substrate, for example of silicon, having a diameter of 200 mm, with a resistivity p preferably of between 1 and 50 ohm / cm and having alignment marks advantageously on the front face and cavities 9, extending over the entire first face 7 of the second substrate 5 as in Figure 3, is cleaned with a solution such as an ozonated water solution having 40 mg / l of ozone, with a solution of APM (ammonium peroxide mixture) with a concentration of ammonia, hydrogen peroxide and deionized water respectively 0.25 / 1/5 in volume proportions. A second substrate 5, for example size, material and identical marks is also used and cleaned. Then the two substrates 1, 5 are dried in a dryer.

 In general, the formation of the intermediate layer of the invention can be carried out in different ways. One or more drops can be deposited firstly on one of the substrates. The drop or drops are then spread. According to a possibility applicable to the example given here, to form the interlayer water film, one of the substrates, preferably without cavities, is wet and a fast rotational movement makes it possible to obtain a liquid film of fine thickness, by example 100nm. Spreading can also be produced or perfected by the contact of the other substrate, applying on the other side of the film to be formed.

Then the substrates are brought together until a liquid mattress having a thickness of, for example, 100 nm is obtained between the substrates 1, 5 as in FIG. 4. This approximation can be completely mechanical or simply by dropping the second substrate 5 from above. on the first. It can advantageously be a combination of the two to spread the drop well from the center of the first face 3 towards the edges and then let gravitation finish the approximation by using the liquid mattress that is formed to ensure the physical separation of the substrates 1, 5 before the end of the alignment. This alignment can take place during the reconciliation and after the latter or only once the substrates are very close to one another. The alignment control can be done by infrared. Once the alignment is completed, the vacuum is advantageously made around the plates to remove the material from the interlayer. The open cavities 9 allow faster removal of this material. As long as the material is not removed and the direct bonding has not started, the alignment is checked and corrected if necessary. As soon as direct bonding is triggered, alignment control is no longer necessary, as in figure 5. On the other hand, it is desirable to keep the plates under vacuum until the direct bonding is effective on the 3. A variant may be to replace the vacuum with an anhydrous nitrogen atmosphere having a dew point temperature of less than -10 ° C or less than -85 ° C.

 The present invention is not limited to the previously described embodiments but extends to any embodiment within its spirit.

Claims

1. A method of assembling a first face (3) of a first substrate (1) and a first face (7) of a second substrate (5), characterized in that it comprises:  a formation of an interlayer (1 1) between the first face (3) of the first substrate (1) and the first face (3) of the second substrate (5), the intermediate layer (1 1) being in contact with the first face (3) of the first substrate (1) and the first face (7) of the second substrate (5);  - Elimination of the intermediate layer (1 1) causing a contacting of the first face (3) of the first substrate (1) and the first face (7) of the second substrate (5).  2. The assembly method according to the preceding claim, wherein a check is made of the alignment of the first and second substrates (1, 5) before and / or during the removal of the interlayer (1 1).  3. Assembly method according to the preceding claim, wherein the control of the alignment comprises a change in the relative position of the first and second substrates in a plane parallel to the first faces of the first and second substrates. 4. Assembly method according to one of the preceding claims, wherein the removal of the intermediate layer (1 1) comprises an evaporation of the material of said intermediate layer (1 1).  5. Assembly method according to one of the preceding claims, wherein the elimination of the intermediate layer (1 1) comprises a discharge of the material of said intermediate layer (1 1) outside the first faces (3,7) first and second substrates (1, 5).  6. Assembly method according to the preceding claim, wherein the evacuation comprises the formation of a vacuum around the first and second substrates (1, 5). 7. The method of assembly according to one of the two preceding claims, wherein the evacuation is operated at least in part via at least one cavity (9) opening on the surface of the first face of at least one of the first and second substrates (1, 5). 8. Assembly method according to the preceding claim, wherein the at least one cavity (9) is formed to open on a wafer portion of the substrate (1, 5).  9. The assembly method according to one of the preceding claims, wherein the step of contacting is configured to produce direct bonding between the first faces (3,7).  10. The assembly method according to one of the preceding claims, wherein the material of the intermediate layer is incompressible.  1 1. Assembly method according to one of the preceding claims, wherein the material of the intermediate layer (1 1) is a liquid.  12. The assembly method according to the preceding claim, wherein the material of the intermediate layer (1 1) has a viscosity less than 1 pascal. second at 20 ° C.  13. The assembly method according to one of the preceding claims, wherein the material of the interlayer (1 1) is selected from water, a solvent including acetone, an alcohol.  14. The assembly method according to one of the preceding claims, wherein the intermediate layer (1 1) is chosen with a thickness less than 100 nm. 15. Assembly method according to one of the preceding claims, wherein the formation of the intermediate layer (1 1) comprises a deposit of at least one drop (10) of a material intended to form the intermediate layer (1). 1) on the first substrate (1).  16. The assembly method according to the preceding claim, wherein a drop (10) is deposited in the center of the first face (3) of the first substrate. (1) - 17. The assembly method according to one of the two preceding claims wherein the formation of the intermediate layer (1 1) comprises a spread of the at least one drop by rotating the first substrate (1) along a directed axis. according to its thickness. 18. Method according to one of the three preceding claims, wherein the formation of the intermediate layer comprises a relative approximation of the first and second substrates (1, 5) configured to spread the at least one drop (10) on the first faces (3,7) of the first and second substrates (1,5).  19. The assembly method according to the preceding claim, wherein the relative approximation is operated on at least a final portion of its stroke by gravitational descent of the second substrate (5) to the first substrate (1). 20. Method according to the preceding claim, wherein the gravitational descent is operated by releasing the second substrate (5) of the right of way of a holding device.  21. Method according to the preceding claim, wherein, after the relative approach, reactivates the grip of the holding device on the second substrate (5).  22. Assembly method according to one of the preceding claims, wherein forming at least one microelectronic component on at least one of the first substrate (1) and the second substrate (5). 23. A method of assembly according to the preceding claim, wherein forming a first microelectronic component on the first substrate (1) and a second component on the second substrate (5), contacting the first face (3) of the first substrate (1) and the first face (3) of the second substrate (5) being configured to contact the first and second microelectronic components.